Re: [Discuss-gnuradio] Agile Solutions is Pleased to announce USRP STAR. Yet another addition to GNURADIO

[info]

On Sun, 17 Apr 2011 23:09:15 -0700, Matt Ettus <address@hidden> wrote:
> On 04/16/2011 11:15 AM, address@hidden wrote:
>> Hi Marcus,
>>
>>
>>
>> Welcome to Agile Solutions.
>>
>>
>>
>> Yes, it uses both common clock and common LO.  Theres no phase tracking
>> required due to common LO with cycle slip. This is main requirement for
>> MIMO systems for Space Time(STBC) and Space Frequency(SFBC) Codes.
>>
>>
>>
>> Although, diversity only systems needs only common clock. Either TX or
>> RX. There were fundamental issues when we used USRP1 to detect this
>> problem. We had extensive analysis and solid theoretical analysis to
>> prove this. Although this analysis and material is not property of Agile
>> Solutions, we would not like to comment more on this.
> 
> 
> Dear Mr. Agile Solutions,
> 
> You appear to have a fundamental misunderstanding of MIMO systems in
> general, and of the design and function of the USRP in particular.  The
> USRP1 (as well as all of our other products) is fully MIMO capable.  It
> does not suffer from cycle slips, and all oscillators are fully coherent
> so the phases "track".
> 
> Many papers have been written by people using USRPs for MIMO, including
> with STBC and SFBC, as well as phased arrays.  For example, a 30-second
> search on Google finds this one right away:
> 
>       http://newport.eecs.uci.edu/~hyousefi/publ/sdrMC08.pdf
> 
> Matt Ettus
> President, Ettus Research LLC


Dear Matt,

We do not have any fundamental misunderstanding of MIMO systems.
Further, there are fundamental drawbacks using common clock for multiple
LO's used for achieving phase coherency. 

Should you need for more understandings of MIMO systems, below are the
leading Test and Measurement companies. Please check this post from 
NI http://zone.ni.com/devzone/cda/pub/p/id/1001 
and AGILENT
http://wireless.agilent.com/wireless/helpfiles/n7617b/n7617b_technical_overview.pdf

USRP STAR meets all the design requirements for a true MIMO system. A
host of MIMO transmitter and receiver examples we will be posting soon.
These will be plug and play. People can use these examples to build
their system. 

We went through this published article you sent to us for review
http://newport.eecs.uci.edu/~hyousefi/publ/sdrMC08.pdf

We didn't see anything particular to STBC/SFBC. For a beamformong
system phase coherence is the fundamental requirement. By varying
phase(weighting the antenna co-efficient) we can steer the beam. In
USRP1/2/200/210 there is no guarantee of the phase reference every time
you send the packets for transmitting or receiving. Every packet
transmission they start with arbitrary phase. So every time the beam
will be switching in arbitrarily. Particularly transmitter DAC doesn't
have control over the phase plus deviation in the LO phase lock with
track length of the common clock. And for for a phased array system,
DOA(Direction of arrival) depends on the phase reference of the system.
So every time the phase reference needs o be tracked in the algorithm to
know actual DOA. For a diversity transmitter and diversity receiver
employing MRC would only be realized with USRP1 system as per our
understanding.

We would like to know if USRP1 is capable of these things and how ?    
 
Although we haven't come across any projects with USRP1 for successful
testing for STBC and SFBC codes. 
If any team or individual has published this work, we would be happy to
check that part.

To verify from our end we took IQ samples for 2x2 Almouti scheme from
Agilent-VSA example files. 
We have then transmitted and received these IQ samples via
USRP1+RFX2400. Transmission and reception done at 1MSPS with 4x
interpolation.
Received IQ samples when loaded into VSA for analysis, it failed to
demodulate and decode. We have used Agilent VSA for years now and we
consider it as a golden reference for demodulation and decoding. 

We respect your timely suggestions and advice.

Thank you.